Network centric workload in data centers often requires identical processing by several proxies and devices in a network pipeline, for example, before the workload is processed by a front-end application or back-end server. For example, processing of protocol requires maintaining states in a computer network system, which can cause significant processing latency, and doing so at several devices would compound the latency. A mechanism that could reduce a processing latency in processing network centric workload may be desirable.